Multitask LQG Control: Performance and Generalization Bounds
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910142893654016 |
|---|---|
| author | Toso, Leonardo F. Fallah, Kasra Stamouli, Charis Pappas, George J. Anderson, James |
| author_facet | Toso, Leonardo F. Fallah, Kasra Stamouli, Charis Pappas, George J. Anderson, James |
| contents | We study multitask learning for stochastic and partially observed control systems, focusing on the linear quadratic Gaussian (LQG) problem. Our goal is to learn a common stabilizing controller that generalizes across a distribution of systems and objectives. To this end, we leverage a history-dependent lifting that recasts the multitask LQG problem into an equivalent high-dimensional multitask LQR problem, allowing for the analysis of policy gradient methods. We show that learning a common lifted controller induces a heterogeneity bias which we characterize via a "bisimulation function". We establish performance and generalization guarantees that explicitly depend on such bisimulation-based heterogeneity measures. For model-free, we demonstrate that multitask learning reduces policy gradient estimation variance proportionally to the number of tasks in the training set. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_16730 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Multitask LQG Control: Performance and Generalization Bounds Toso, Leonardo F. Fallah, Kasra Stamouli, Charis Pappas, George J. Anderson, James Optimization and Control We study multitask learning for stochastic and partially observed control systems, focusing on the linear quadratic Gaussian (LQG) problem. Our goal is to learn a common stabilizing controller that generalizes across a distribution of systems and objectives. To this end, we leverage a history-dependent lifting that recasts the multitask LQG problem into an equivalent high-dimensional multitask LQR problem, allowing for the analysis of policy gradient methods. We show that learning a common lifted controller induces a heterogeneity bias which we characterize via a "bisimulation function". We establish performance and generalization guarantees that explicitly depend on such bisimulation-based heterogeneity measures. For model-free, we demonstrate that multitask learning reduces policy gradient estimation variance proportionally to the number of tasks in the training set. |
| title | Multitask LQG Control: Performance and Generalization Bounds |
| topic | Optimization and Control |
| url | https://arxiv.org/abs/2604.16730 |